besselj bug @cygwin32

[John W. Eaton]

On 19-Feb-1998, Rolf Fabian <address@hidden> wrote:

| BUG-REPORT for platform
| cygwin32-octave2.0.10-binary downloaded 980217 from your server)
| 
|    besselj(N,x) - function, now implemented in 2.0.10
|    produces OVERALL GARBAGE even for the most simple 'besselj(0,x)'

Yikes!  You are quite right.  The patch below should fix the problem.
I will be making another release in a few days to fix this and a few
other problems with 2.0.10.

Thanks,

jwe


--- libcruft/specfun/ribesl.f~  Sun Nov 30 03:37:59 1997
+++ libcruft/specfun/ribesl.f   Thu Feb 19 20:25:31 1998
@@ -169,7 +169,7 @@
 C-------------------------------------------------------------------
       LOGICAL FIRST
       INTEGER IZE,K,L,MAGX,N,NB,NBMX,NCALC,NEND,NSIG,NSTART
-      DOUBLE PRECISION DGAMMA, D1MACH
+      DOUBLE PRECISION DGAMMA, D1MACH,
      1 ALPHA,B,CONST,CONV,EM,EMPAL,EMP2AL,EN,ENMTEN,ENSIG,
      2 ENTEN,EXPARG,FUNC,HALF,HALFX,ONE,P,PLAST,POLD,PSAVE,PSAVEL,
      3 RTNSIG,SUM,TEMPA,TEMPB,TEMPC,TEST,TOVER,TWO,X,XLARGE,ZERO
@@ -192,7 +192,7 @@
 C  Machine-dependent parameters
 C-------------------------------------------------------------------
       IF (FIRST) THEN
-        NSIG = NINT (-LOG (D1MACH (1)))
+        NSIG = NINT (-LOG (D1MACH (4)))
         ENTEN = 1.0D1 ** (INT (LOG10 (D1MACH (2))))
         ENSIG = 1.0D1 ** NSIG
         RTNSIG = 1.0D1 ** (-NINT (NSIG / 4.0))

--- libcruft/specfun/rjbesl.f~  Sun Nov 30 03:37:59 1997
+++ libcruft/specfun/rjbesl.f   Thu Feb 19 20:47:16 1998
@@ -139,7 +139,7 @@
 C---------------------------------------------------------------------
       LOGICAL FIRST
       INTEGER I,J,K,L,M,MAGX,N,NB,NBMX,NCALC,NEND,NSIG,NSTART
-      DOUBLE PRECISION  DGAMMA, D1MACH
+      DOUBLE PRECISION  DGAMMA, D1MACH,
      1 ALPHA,ALPEM,ALP2EM,B,CAPP,CAPQ,CONV,EIGHTH,EM,EN,ENMTEN,ENSIG,
      2 ENTEN,FACT,FOUR,FUNC,GNU,HALF,HALFX,ONE,ONE30,P,PI2,PLAST,
      3 POLD,PSAVE,PSAVEL,RTNSIG,S,SUM,T,T1,TEMPA,TEMPB,TEMPC,TEST,
@@ -183,7 +183,7 @@
 C  Machine-dependent parameters
 C---------------------------------------------------------------------
       IF (FIRST) THEN
-        NSIG = NINT (-LOG (D1MACH (1)))
+        NSIG = NINT (-LOG (D1MACH (4)))
         ENTEN = 1.0D1 ** (INT (LOG10 (D1MACH (2))))
         ENSIG = 1.0D1 ** NSIG
         RTNSIG = 1.0D1 ** (-NINT (NSIG / 4.0))